Add forests and high column density intervening absorbers

[matpieri]

In addition to QSO frame templates we wish to add intervening absorbers to generate low-level mocks for redshift and identification and null bricks with no cosmological cross-correlations.

The diversity of QSO templates (provided by Isabelle Paris) will be expanded to including the diversity of forest realisation plus the diversity of high-column density lines (LLS, sub-DLA and DLA). LLS will have negligible impact here, but sub-DLA and DLA may have a significant noise. Intrinsic DLAs may also be problematic for QSO redshirting if the Lya emission line is significantly suppressed.

We plan provide suite of mock skewers randomly drawn from the BOSS mock catalogues for various high-z redshift intervals (led by Michael Blomqvist). There should be a sufficient number to Monte Carlo the diversity of QSO templates at every redshift. We will also ensure that HCDs are included with a sufficient level of accuracy and allow for a sufficient sample size to cover the diversity of HCDs (led Debopam Som).

The QSO templates can be combined with the suite of mock skewers minimally to test redshift and id, or scale to produce bricks

[orionsom]

This week we will use the high column density lines (including LLS, sub-DLA and DLA over range in column from N_HI>17.2) as currently present in BOSS mocks. This implementation should be sufficient for the first pass but can be improved in the following aspects:

1. Using the observed sub-DLA (log NHI = 19-20.3) differential column desnity distribution function (Zafar et al. 2013; http://www.aanda.org/articles/aa/pdf/2013/08/aa21154-13.pdf) instead of the model predictions (Figure 1, Font-Ribera & Miralda-Escude 2012; http://iopscience.iop.org/article/10.1088/1475-7516/2012/07/028/pdf) incorporated currently.
2. Using updated observed DLA differential column density distribution function, available from Noterdaeme et al. 2012 (http://www.aanda.org/articles/aa/pdf/2012/11/aa20259-12.pdf).

[dkirkby]

Thanks for the update Debopam. Even something very simple would be useful to get us started, for example a single DLA rest-frame absorption function that we can use to multiply the flux in randomly selected simulated quasars.

[orionsom]

Hi, I am attaching normalized Voigt profiles (2 column ascii files) for 3 sub-DLA/DLAs (log NHI of 19.5, 20.5 and 21.5; b=25 km/s). These profiles are in velocity space (del_v ~25 km/s). Hopefully these will be okay to get things going. I shall try to send something more refined tomorrow. lgnhI19.5.txt lognhI20.5.txt lognhI21.5.txt

[matpieri]

Great that you did this too Debopam. Duplicating effort here but in this folder https://www.dropbox.com/sh/tu2w3a0qax71faj/AAAUpHz-_CKQ2WnOOowVmyFOa?dl=0 find a series of 8 columns ranging from 10^19 to 10^22.5 In each file the 1st column is absorber restframe wavelength the 4th column is the profile. Ignore everything else.

Note that the next step is to provide the BOSS mock forest skewers with the DLAs optionally in or out depending on the column used. These will be the raw mock skewers before MockExpander has done its work. These are on a (slightly) finer grid than SDSS - 50.7 km/s intervals. We plan to offer them up in a convenient fashion with the wavelength solution clearly stated and tabulated by quasar redshift. As I said earlier this will be a suite of skewers to draw from with DLAs included.

We chose this on DLAs because the DLA placement in the BOSS mock LSS are better than something we can put together this week. After some checks the DLA frequency seem fine for now, with only the sub-DLA number density per dz in need of serious attention as Debopam said.

On 3 Nov 2015, at 23:58, Debopam Som notifications@github.com wrote:

Hi, I am attaching normalized Voigt profiles (2 column ascii files) for 3 sub-DLA/DLAs (log NHI of 19.5, 20.5 and 21.5; b=25 km/s). These profiles are in velocity space (del_v ~25 km/s). Hopefully these will be okay to get things going. I shall try to send something more refined tomorrow. lgnhI19.5.txt lognhI20.5.txt lognhI21.5.txt

— Reply to this email directly or view it on GitHub.

Matthew Pieri A*MIDEX Excellence Chair Laboratoire d'Astrophysique de Marseille Aix-Marseille University 38, rue Frédéric Joliot-Curie 13388 Marseille cedex 13 FRANCE

matthew.pieri@lam.fr +33-(0)4-91-05-69-87

[dkirkby]

@orionsom Just to make sure I understand how to use these tables: if I read one of your files into vectors dv and voigt, and I have a QSO spectrum, wlen and flux, tabulated in the observed frame, then to imprint the absorption feature, I calculate:

new_flux = interpolated_voigt * flux

where interpolated_voigt is interpolated to

dv = clight * (wlen - wlen_abs) / (wlen_lya)

and wlen_abs < wlen_qso is the absorber redshift. What is a reasonable probability density for wlen_abs ?

[orionsom]

@dkirkby I think the correct mapping from vector dv to vector wlen should be represented by dv = clight * (wlen - wlen_abs) / (wlen_abs) ? when wlen_abs < wlen_qso

[matpieri]

and wlen_abs < wlen_qso is the absorber redshift. What is a reasonable probability density for wlen_abs ?

We are minutes away from producing the mock skewers with DLAs and sub-DLAs present following the (broadly) right probability density. At this point we’d like to focus on the use of those. The DLA profiles we have provided can be used if you want to do case studies putting 1 DLA in by hand.

[dkirkby]

@orionsom Yes!

@matpieri We still want to be able to do studies with, eg., the same template w/ and w/o a "fiducial DLA", etc.

[matpieri]

@dkirkby

You will be able to. Our mocks with forests will be available in two forms with DLA and without.

On 04 Nov 2015, at 16:09, dkirkby notifications@github.com wrote:

@orionsom https://github.com/orionsom Yes!

@matpieri https://github.com/matpieri We still want to be able to do studies with, eg., the same template w/ and w/o a "fiducial DLA", etc.

— Reply to this email directly or view it on GitHub https://github.com/dkirkby/ArgonneLymanAlpha/issues/8#issuecomment-153756370.

[matpieri]

@dkirkby Its is possible to change this issue title to include be “Add forests and high column density absorbers to generate low-level mocks” or similar? I realise now that I mislabelled the issue.

[matpieri]

For tracking purposes in this issue. As states elsewhere

Michael and I are working on mock skewers with high column density lines included, which we will share. I will then generate an astrophysical mock sample and fold in Isabelle's QSO templates,which we will also share.

Debopam is learning how to run specsim to turn these astrophysical mocks to turn them into low-level DESI mocks.

On 04 Nov 2015, at 16:09, dkirkby notifications@github.com wrote:

@orionsom https://github.com/orionsom Yes!

@matpieri https://github.com/matpieri We still want to be able to do studies with, eg., the same template w/ and w/o a "fiducial DLA", etc.

— Reply to this email directly or view it on GitHub https://github.com/dkirkby/ArgonneLymanAlpha/issues/8#issuecomment-153756370.

[dkirkby]

@debopam Look at simulate/qso_grid.py in this repo.

[dkirkby]

Is there any plot I should include in our update (9 mins from now) ?

[matpieri]

Not right now I'm afraid. documentation coming your way. We would like to be able to combine a forest + template, but we probably can't do that in time.

On 4 Nov 2015, at 17:51, dkirkby notifications@github.com wrote:

Is there any plot I should include in our update (9 mins from now) ?

— Reply to this email directly or view it on GitHub.

Matthew Pieri A*MIDEX Excellence Chair Laboratoire d'Astrophysique de Marseille Aix-Marseille University 38, rue Frédéric Joliot-Curie 13388 Marseille cedex 13 FRANCE

matthew.pieri@lam.fr +33-(0)4-91-05-69-87

[matpieri]

5000 mock forest + HCD lines (LLS + sub-DLA +DLA) are now available here https://www.dropbox.com/sh/y2dnw0n5so04xq5/AADt07yiR3gvuUsXXNG0FnLQa?dl=0

They are randomly drawn from the whole BOSS mocks catalogue http://adsabs.harvard.edu/abs/2015JCAP...05..060B such such that cross-correlations should be absent. No two mock skewers are used twice.

5000 mocks forests without DLAs are in skewers_no_dla.fits The same 5000 mocks forests this time with HCD are in skewers_dla.fits The fits files contain 5000 columns, 1 per spectrum and no header.

Each column corresponds two 1 line in random_skewers_info_M3_0_0.txt which shows [ column/spectrum number] [original file name] [random realisation used (from 10)] [QSO z]

They are on a common wavelength grid stated in wavegrid.txt

The idl code used to create this is in the folder.